I would like to add ADC Overflow indicator with hold (approx 5sec) to AK5385A chip but I'm not sure about circuit to go with :/ any recomendations? 
AK5385A ADC Overflow indicator
- Thread starter Moby
- Start date
Help Support GroupDIY Audio Forum:
Diode into a resistor/cap (in parallel) to ground, followed by a buffer.
Google some "peak hold" circuits. The path for the cap to discharge should be through the resistor to ground. That will set up your time constant. Input to buffer should be ultra hiz, otherwise current from the cap will discharge through the buffer.
Google some "peak hold" circuits. The path for the cap to discharge should be through the resistor to ground. That will set up your time constant. Input to buffer should be ultra hiz, otherwise current from the cap will discharge through the buffer.
thanks, I will
How about this? I'm not sure can it "hold" the peak :/
http://electrosuite.com/audio/audio-clipping-indicator.html#axzz2G9yUepHk
http://electrosuite.com/audio/audio-clipping-indicator.html#axzz2G9yUepHk
Theres no need for an extra analog detector. The adc already has an overflow pin!
Take the output of the pin through a diode, and into a parallel r and c (both connected to ground on one side.. Take the signal side into high impedance buffer, such as a fet or opamp.
The decay time will be defined by the vale of r and c.
Take the output of the pin through a diode, and into a parallel r and c (both connected to ground on one side.. Take the signal side into high impedance buffer, such as a fet or opamp.
The decay time will be defined by the vale of r and c.
Wow, yes, I was blind when I was reading AK5385A datasheet :/ definately pin 9 "Analog Input Overflow Detect PinRochey said:
This pin goes to “H” if analog input overflows"
I wonder is it possible to use monostable multivibrator to achieve LED on hold. Something like this?
It is probably easier to make a simple hold circuit with cap and buffer. No need for a precise hold time.
JR
JR
Thanks, do you have some circuit example?JohnRoberts said:
This seems pretty simple... Rochey already suggested a simple circuit.
To break it down you need to convert a brief logic 1 (?) from the IC, to a longer duration current fed through a LED .
Starting with Rochey's suggestion to
Anode end of simple signal diode connects to the overflow pin.
Cathode (banded) end of diode connects to both a capacitor and resistor.
Other end of capacitor connects to ground.
The other end the resistor could be connected to ground.
When the overflow flag momentarily goes high, it will charge up the cap relatively quickly. When the overflow flag goes low the cap will slowly discharge at the RxC rate. You could almost connect a high efficiency LED directly to the cap through a current limit resistor, but that will not give much hold time and load the logic output
A very simple current buffer could connect the ground end of the discharge resistor to the base of a general purpose NPN transistor instead. Connect the emitter of that NPN transistor to ground. Connect the LED with a current limit resistor in series between a +V supply and the NPN collector. Connect the LED anode to +V, the LED cathode to the current limit resistor then collector.
Selecting values for this will be good mental exercise, but the circuit should work with a fairly wide range of values.
Good luck...
JR
To break it down you need to convert a brief logic 1 (?) from the IC, to a longer duration current fed through a LED .
Starting with Rochey's suggestion to
Rochey said said:
Anode end of simple signal diode connects to the overflow pin.
Cathode (banded) end of diode connects to both a capacitor and resistor.
Other end of capacitor connects to ground.
The other end the resistor could be connected to ground.
When the overflow flag momentarily goes high, it will charge up the cap relatively quickly. When the overflow flag goes low the cap will slowly discharge at the RxC rate. You could almost connect a high efficiency LED directly to the cap through a current limit resistor, but that will not give much hold time and load the logic output
A very simple current buffer could connect the ground end of the discharge resistor to the base of a general purpose NPN transistor instead. Connect the emitter of that NPN transistor to ground. Connect the LED with a current limit resistor in series between a +V supply and the NPN collector. Connect the LED anode to +V, the LED cathode to the current limit resistor then collector.
Selecting values for this will be good mental exercise, but the circuit should work with a fairly wide range of values.
Good luck...
JR
Very detailed Thanks John
Thanks John
I experimented a bit with this, but thing I really don't like is the LED light fade. I was afraid of that but I was not lazy to try it. What can I do about it?JohnRoberts said:
HysteresisMoby said:
JR
Feature creep, like adding snap action hysteresis requires more parts and makes the consulting bill more expensive.
Of course one simple thing to try is to add one more resistor from the base to ground, that should turn it off a little sharper.
JR
Of course one simple thing to try is to add one more resistor from the base to ground, that should turn it off a little sharper.
JR
I tried already but not a big improvement. So, Monostable Multivibrator is the way to go . Not 555 because it needs negative impulse, maybe SN74LS02 or so...JohnRoberts said:
Google up the old school two transistor flip-flop. IIRC It's called an "astable multi-vibrator" or something like that. Simply uses hysteresis or positive feedback to snap into one mode, then snap back to steady state mode.
A 555 timer is easy but not the same educational experience. Once upon a time transistor FF were how it was done, before the rise of ICs.
JR
A 555 timer is easy but not the same educational experience. Once upon a time transistor FF were how it was done, before the rise of ICs.
JR
Wouldn't a comparator do the trick?
Rochey said:
There are way too many ways to accomplish this to count.
Yes something like a 339 comparator with + and - inputs could feedback some of the output state change back into an input to create threshold hysteresis (shift) for snap action output state change.
While if you truly want to KISS, don't dismiss my suggestion to add a resistor from base to ground. You might also need to scale other associated values somewhat.
The issue with the transistor turning off too slow is because it is operating in base current amplification mode to multiply that base current to drive the collector load. Instead driving the base with a voltage (fraction of the cap voltage divided down) will cause the current to drop more quickly (exponentially) between full on at say .7 and full off at say .4V.
In base current amplification mode, the base current is the cap voltage - the base voltage (say 5 - .5V), divided by the series R.
In base voltage drive mode, a resistor divider divides down the cap voltage between way above .5V for on, to just below .5V for off at the base.
You may want to use a larger cap and smaller R values for the voltage mode to be more voltage like to the transistor.
=====
Pretty much any discrete circuit is as simple or complex as you are willing to analyze it.
JR
Similar threads
Latest posts
-
-
-
-
-
-
-
-
EZ1084EQ NEW BATCH COMING - SEE MY MARCH 27 COMMENT ON PAGE 38
- Latest: sodiumthiosulphate